Population-based Training (PBT) primarily falls into the categories of Artificial Intelligence, Big Data and Smart Data, and Digital Society. This term refers to a method for improving artificial intelligence (AI) by basing the training of algorithms on a wide range of data from the entire population, rather than just a small, selected dataset.
Imagine an AI is intended to assist with medical diagnoses. If it is only trained on data from a small group of people, it could produce inaccurate results for other population groups. In population-based training (PBT), algorithms are trained with very large and diverse datasets – for example, with anonymised health data from millions of people of different backgrounds, age groups, and regions.
This increases the accuracy and reliability of AI models in everyday life. So, PBT ensures that digital applications work better for everyone – whether it's personalised medicine, traffic control, or smart household appliances. This makes artificial intelligence fairer and more widely usable, as it is based on genuine data diversity from society.
Deutsch 
English (UK)